By tapping into artificial intelligence-based product recommendations, the sportswear merchant can better understand what its customers are shopping for and show them a relevant product as they browse throughout the ecommerce site.

MandM Direct Ltd., an online-only sportswear retailer with more than 15,000 products for sale, found that the product recommendations around its ecommerce site were lacking. The online merchant made the switch from a rules-based product recommendation strategy to an artificial intelligence-based system about a year ago.

“We want to understand every action a user does, especially what will make them do the behavior that is most valuable to the business,” says Paul Allen, head of ecommerce. “If the customer is adding another product to their basket or getting through checkout, that adds value to the business. And to do that, we want to put relevant content in front of them.”

By using a rules-based recommendation system, MandM Direct was missing out on valuable customer data. The recommendations showed shoppers products based on categories they had browsed, the products they’d looked at and their geographic location, but it didn’t dig deeper to recommend items on an individual level.

“Rules-based makes that difficult,” Allen says. “You can maybe hit 50-60% of customers with that strategy. But we wanted to use AI to have more coverage of customers.”


For example, an AI system recommends products that a customer is most likely to want next while in her existing customer journey, he says, such as an item related to their past purchases and the items they’d added to their basket.

The U.K.-based merchant has worked with personalization software company Qubit for more than seven years on its ecommerce site for personalization and its rules-based recommendations, so it tapped the vendor to help with integrating AI recommendations. MandM Direct began talks to integrate AI in February 2020, then began testing it in April. It tested the feature in two rounds and completed its testing by October before launching it sitewide.

“Our main challenge was really getting the tests set up and testing two different types of AI recommendations at the same time,” Allen says. “But after that point, it’s been pretty smooth sailing ever since. We’ve just tweaked and improved.”

The new feature helped boost ecommerce sales, even during the testing phase. MandM Direct ended its fiscal 2020 in July with nearly $316 million in revenue, Allen says. That’s a 5.5% year-over-year boost in online sales, according to Digital Commerce 360 estimates.


AI product recommendations are used on the homepage, product listing pages (which show products listed in a category or in search results), product detail pages, the shopping cart, search—anywhere a customer is looking for a product, they will see personalized recommendations, Allen says.

On the product listing pages, the new AI feature boosted conversion rate by 1.7% year over year (January 2021 over January 2020). In addition, MandM Direct had a 1.8% uptick in revenue per visitor, as well as a 148% lift in click-through rate.

On product detail pages, AI product recommendations boosted conversion rate by 0.93% year over year. And these recommendations lifted revenue per visitor by 0.72% and click-through rate by 5.28%.

On the homepage, MandM Direct’s conversion rate grew by 0.6% year over year thanks to the AI product recommendations. And its revenue per visitor increased 0.7% and its click-through rate jumped 118%.

Product recommendations on the add-to-bag page boosted click-through rate by 157% year over year.

“AI product recommendations come up with customer insights we couldn’t get with traditional analytics,” Allen says. “If you can get that in front of the customer at the right time, they are more likely to make that purchase.”

The retailer declined to reveal the cost of the AI recommendation engine, but Allen says  it takes MandM Direct’s team fewer recourses to manage the new system compared with rules-based recommendations.


In the coming months, Allen says his team is in the early stages of expanding customer data sources with in-session data that will enable MandM Direct to define and understand the customers’ “mission” each time they visit to provide personalization in near real time while they’re browsing.

“It’s all about getting the right product in front of the customer at the right time,” Allen says.

MandM Direct is No. 137 in the Digital Commerce 360 Europe 500.